Retraining and Updating Knowledge
Retrain your Data Sources to refresh AI knowledge after content changes.
When to retrain
Retrain when:
- Website content changes
- Pages are removed
- Documents are updated
- Pricing or policies change
Outdated content results in outdated answers.
Full retrain vs selective retrain
| Option | Description |
|---|---|
| Full retrain: | Reprocesses entire source. Recommended after major changes |
| Selective retrain: | Use when updating specific pages or files. If many pages changed, use full retrain. |
Website updates
After updating your website:
| Step | Action |
|---|---|
| 1 | Go to Dashboard -> Data Sources |
| 2 | Select your website |
| 3 | Click Retrain |
Manual re-scrape
Use manual re-scrape when you need to refresh website content immediately.
| Step | Action |
|---|---|
| 1 | Go to Dashboard -> Data Sources |
| 2 | Select your website source |
| 3 | Click Re-scrape |
Manual re-scrape checks for modified content and newly added pages before updating scraped data.
Auto re-scrape intervals
Auto re-scrape can refresh website sources on an interval.
| Interval |
|---|
| Every 7 days |
| Every 15 days |
| Every 30 days |
Auto re-scraping checks for modified content and newly added pages before updating scraped data.
Handling large content changes
For large updates:
- Remove outdated sources first
- Confirm new content is live
- Run full retrain
This prevents duplicate or conflicting information.
Important Notes or Limitations
Training time depends on content size and plan limits.